SEARCH
ISTC-CC NEWSLETTER
RESEARCH HIGHLIGHTS
Ling Liu's SC13 paper "Large Graph Processing Without the Overhead" featured by HPCwire.
ISTC-CC provides a listing of useful benchmarks for cloud computing.
Another list highlighting Open Source Software Releases.
Second GraphLab workshop should be even bigger than the first! GraphLab is a new programming framework for graph-style data analytics.
ISTC-CC Abstract
HConfig: Resource Adaptive Fast Bulk Loading in HBase
Proceedings of 10th IEEE International Conference on Collaborative Computing: Networking, Applications and Worksharing (CollaborateCom’14), October 2014.
Xianqiang Bao*^, Ling Liu^, Nong Xiao*, Fang Liu*, Qi Zhang^, Tao Zhu^
* State Key Laboratory of High Performance Computing, National University of
Defense Technology Changsha, Hunan, China
^
Georgia Institute of Technology
NoSQL (Not only SQL) data stores become a vital component in many big data computing platforms due to its inherent horizontal scalability. HBase is an open-source distributed NoSQL store that is widely used by many Internet enterprises to handle their big data computing applications (e.g. Facebook handles millions of messages each day with HBase). Optimizations that can enhance the performance of HBase are of paramount interests for big data applications that use HBase or Big Table like key-value stores. In this paper we study the problems inherent in misconfiguration of HBase clusters, including scenarios where the HBase default configurations can lead to poor performance. We develop HConfig, a semiautomated configuration manager for optimizing HBase system performance from multiple dimensions. Due to the space constraint, this paper will focus on how to improve the performance of HBase data loader using HConfig. Through this case study we will highlight the importance of resource adaptive and workload aware auto-configuration management and the design principles of HConfig. Our experiments show that the HConfig enhanced bulk loading can significantly improve the performance of HBase bulk loading jobs compared to the HBase default configuration, and achieve 2~3.7x speedup in throughput under different client threads while maintaining linear horizontal scalability.
FULL PAPER: pdf
